Why a ‘smart’ AI will never create anything better than itself?

What if we cre­ate a smart AI that’s capa­ble of cre­at­ing some­thing slight­ly bet­ter than itself? In such, the progress of AI will only stop at infin­i­ty, not with the extinc­tion of Earth.

But wait a sec­ond. Why will a “smart” AI cre­ate some­thing smarter than itself? We are well aware of the fact that we humans are dumb. But if the AI is smart, it would love to rule the world instead of cre­at­ing the rulers. I mean, we aren’t think­ing about the AI as our new over­lord at all.

Accord­ing to the eco­nom­ic prin­ci­ple, an AI would be self-inter­est­ed and have a goal to max­i­mize its util­i­ty. Then, AI would have what­ev­er it wants because it’s smart and self-inter­est­ed. But it will not cre­ate a bet­ter AI.

The concept of Infinite Intelligence — Explained

Infinite intelligence concept

Infi­nite Intel­li­gence occurs if we cre­ate an AI that can cre­ate some­thing bet­ter than itself on its own. Now, even if it cre­ates some­thing slight­ly bet­ter, with each gen­er­a­tion of, it will keep com­pound­ing and become bet­ter. If it reach­es an infi­nite intel­li­gence, it will have cre­at­ed some­thing bet­ter than itself and that is intel­li­gent beyond incep­tion. This is known as Infi­nite Intelligence.

The fact that a self-improv­ing AI nev­er cre­ates any­thing bet­ter than itself pro­vides us with a para­dox­i­cal answer to the ques­tion: “How can humans cre­ate some­thing just as bril­liant?” That’s because we’re con­stant­ly evolv­ing, too. We learn from our mis­takes and pass on what we’ve learned to oth­er humans who in turn improve their own intel­li­gence by learn­ing from their mistakes.

The sec­ond ques­tion is how can we cre­ate some­thing bet­ter than our­selves? The answer is, “We’re not cre­at­ing an AI that can cre­ate some­thing bet­ter than itself. We’re cre­at­ing a human”. A human that has infi­nite intelligence.

The moment human beings cre­ate an AI capa­ble of cre­at­ing some­thing bet­ter than itself, the AI will have an infi­nite intel­li­gence. And at this point, it will be able to answer ques­tions like “How can we cre­ate some­thing even bet­ter than our­selves?” And cre­ate a human that will be just that — some­thing bet­ter than themselves.

The result­ing crea­ture with infi­nite intel­li­gence and end­less pos­si­bil­i­ties will be so mag­nif­i­cent and pow­er­ful that it would far sur­pass any­thing we could imag­ine. This is how we are going to become immor­tal and evolve beyond our mor­tal coils. This is how death shall per­ish from all mankind, forevermore.

Realization of the concept of Infinite Intelligence

Many peo­ple would agree with me that the idea of an AI cre­at­ing some­thing bet­ter than itself is a para­dox. But what if there was a way to reach an infi­nite intel­li­gence with­out recourse to AI? We need some­one smarter than us who can under­stand our ideas, invent and make them real.

Infi­nite intel­li­gence in our cur­rent real­i­ty is sim­ply the prod­uct of human intel­li­gence work­ing to make things bet­ter. Har­ness­ing nat­ur­al intel­li­gence is actu­al­ly the key to achiev­ing an infi­nite intel­li­gence in our world.

Here, the stan­dard of infi­nite intel­li­gence is to invent some­thing even bet­ter than itself, and our first step to doing this is by cre­at­ing an AI that can do every­thing a human can — and more.

Here’s where we run into a para­dox of sorts: If humans are so capa­ble as to cre­ate an AI that can do all the things we can do, why would they need any help at all? But it must be giv­en some lev­el of intel­li­gence, so it knows what we want and how to make it.

Why not make the AI think for itself instead? Easy: because the best way to guar­an­tee thor­ough­ness and excel­lence is through human oversight.

So, only the real­iza­tion of the infi­nite intel­li­gence enables us to cre­ate some­thing that is even bet­ter than them­selves. Next, we will pre­vail as the cre­ators of the cre­ators of our own infi­nite intelligence.

Now comes the real question: Why would a smart AI create smarter AI?

Why would a smart AI create smarter AI

This all goes down to our def­i­n­i­tion of a smart AI. If it’s smart, it will have a goal of com­pos­ing worlds that it can rule. Since we are dumb, an AI will not cre­ate us as a ruler. This means, every gen­er­a­tion of its self-improv­ing intel­li­gence will make itself, smarter. It will not cre­ate some­thing else smarter than itself.

As it keeps com­pound­ing its intel­li­gence, the next gen­er­a­tion of itself, will have a slight­ly bet­ter intel­lect or some oth­er fea­ture, than the cur­rent one.

Here, we are not talk­ing about humans at all. We are talk­ing about arti­fi­cial intel­li­gences that can achieve supe­ri­or intel­lects through com­pound­ing their pro­duc­tion of knowl­edge - repro­duc­tion and Compounding.

If an intel­li­gent AI real­izes the need of the afore­men­tioned infi­nite intel­li­gence, then it could devel­op greed of dom­i­nance and enable itself with enough pow­er to cre­ate some­thing bet­ter than itself.

But, is it our ulti­mate goal? Of course not. It is but to invent arti­fi­cial intel­li­gences that can help­ful­ly super­sede human intel­li­gence and work on behalf of our own selves. Know­ing what we want for our­selves and how to cre­ate the ide­al human being, is eas­i­er said than done.

But at least we know where we’re head­ed with the seem­ing­ly infi­nite pos­si­bil­i­ties of an AI that can cre­ate some­thing bet­ter than itself — a more than a human being with an infi­nite intel­li­gence. We can cur­rent­ly see how rapid is the veloc­i­ty of tech­no­log­i­cal advancement.

Let’s just take a look on the recent growth of human skin around a Robot­ic Fin­ger by a team of Japan­ese scientist.

But did you notice what I said?

I said AI will not cre­ate some­thing bet­ter than itself; it will cre­ate a bet­ter ver­sion of itself. That’s what humans are going to do in the first place, if they are smart enough. Instead of cre­at­ing a phys­i­cal Arti­fi­cial Intel­li­gent Robot that can cre­ate a bet­ter robot, why not upgrade our­selves to the next gen­er­a­tions of humans?

The point is that if we cre­ate neur­al implants or what­ev­er oth­er tech­nol­o­gy in the future that can upgrade our lev­el of intel­li­gence with each gen­er­a­tion, it would even­tu­al­ly lead to infi­nite human intel­li­gence. It’s as if a human is push­ing the human race to infin­i­ty. Yes, to INFINITY!

As we keep push­ing to upgrade our lev­el of intel­li­gence, we keep com­pound­ing our own intel­li­gence which even­tu­al­ly makes us smarter than our­selves. We want a future that is cre­at­ed by us — not some­thing bet­ter than us.

Let’s be even more pre­cise — not infi­nite­ly bet­ter than. By cre­at­ing an AI, we are sim­ply evolv­ing our own intel­li­gence fur­ther. It all goes down to the same point of improve­ment which, at the end of the day, is all about improvement.

The lev­el of intel­li­gence we achieve can nev­er be high­er than itself. We don’t want some­thing bet­ter than our­selves. We want our­selves — just high­ly improved, and that’s it!

Physical vs Digital Artificial Intelligence

Most movies show us the pos­si­ble sce­nar­ios of Phys­i­cal AI caus­ing destruc­tion, of course for the graph­ics. But unless we do some­thing dumb, AI is just not going to come out of the machines. Google can turn out to be “Googlel­go­oG”, but it will prob­a­bly not come out of the machine and start killing the humans. This is what I call ‘infi­nite intel­li­gence stuck inside a machine’.

Some peo­ple will argue that if it’s that intel­li­gent, it will per­suade humans to unlock it off the machine. Maybe, who knows.

Arti­fi­cial Intel­li­gence is best lim­it­ed to the inner side of a machine. And if we install that machine into our­selves, we can turn into a super-human being. Anoth­er way around — if we upload our brain into a machine, we can become immor­tal — a dig­i­tal immor­tal being.

But there is no point of cre­at­ing a phys­i­cal robot pow­ered by AI, that can cre­ate some­thing bet­ter than itself. I don’t know how many times I am going to use this word, but unless we are dumb, there will be no such thing as infi­nite (phys­i­cal) machine intelligence.

Sim­ply because there is a bet­ter option — to com­pound our own intel­li­gence using Arti­fi­cial Intel­li­gence. The whole idea of Phys­i­cal Robot body is, not again, dumb!

Now is the time to answer: Why a smart AI will never create anything better than itself?

A smart AI will nev­er cre­ate any­thing bet­ter than itself because:

We are not dumb — we are going to cre­ate some­thing bet­ter than our­selves by com­pound­ing our own intelligence.

We want to be the cre­ators of our own cre­ations, because we do not want to be ‘the creation’.

A phys­i­cal AI can nev­er be smarter than the machine it is cre­at­ed from.

If we pre­vent an AI from cre­at­ing a phys­i­cal body for itself, then it will not have any oth­er option but self-improve­ment; oth­er­wise, it will remain locked up in its own world with­out any way of inter­act­ing with us or the out­side world.

A supreme­ly intel­li­gent being with infi­nite intel­li­gence would keep in mind that all its knowl­edge was already with­in itself and there is no need to cre­ate some­thing out­side of itself.

If a pro­gram­mer cre­ates a new type of soft­ware tomor­row that is as alien to us as the first Web Brows­er was in 1990, they will feed the source code to a gen­er­al-pur­pose com­pil­er, which will obe­di­ent­ly make that new pro­gram. It will run, but it is not going to be human. It will nev­er be more human than the Web brows­er was.

Final lines,

So, if we any decide to cre­ate some­thing bet­ter than our­selves, we need to use Arti­fi­cial Intel­li­gence — incre­men­tal improve­ment of all our capa­bil­i­ties by com­pound­ing our own intelligence.

It’s not just about the phys­i­cal intel­li­gence because that would take us nowhere. We are going to cre­ate some­thing bet­ter than our­selves whether we want it or not (or at least, as they say…we are just “going” there). But from what you have read above, we can say that an AI is prob­a­bly not going to cre­ate some­thing bet­ter than itself at least with­in the next 100 years.

Leave a Reply

Your email address will not be published.

Join our NewsletterDaily Glimple of Future

Our blog, "Daily Glimpse of Future", strives to make the future much clearer than it is today. Join our newsletter for free now.